AI Detectives Lab: Train Models and Spot Fake Content (Grades 4-6)

Artificial intelligence is shaping how kids search, learn, create, and communicate. Knowing how to ask an AI for an answer is useful, but learners also need to know how AI learns, why it can make mistakes, and how generated content can mislead people. This six-week course uses short visual explanations, games, investigations, and projects so students become AI detectives through experiments rather than long lectures. Students will train and test a simple image-recognition model using Teachable Machine, make predictions, measure accuracy, and iterate to improve results. They will write clear prompts to guide AI ideas, evaluate AI-generated inventions for safety and practicality, and redesign ideas to be more reliable. Learners will investigate suspicious images and online claims by checking sources and evidence, and they will practice verifying information with trusted references. Each class pairs a brief explanation with a guided activity: a game to spot AI vs non-AI, a supervised model-training lab, an invention challenge, a content investigation, a study-guide project using AI, and a final showcase that explores privacy, fairness, and plagiarism scenarios. Generative-AI demonstrations are teacher-supervised; students are not required to create personal generative-AI accounts. No prior AI or coding experience is needed. By the end of the course, learners will be better at spotting misleading content, writing useful prompts, and using AI tools responsibly for learning and creativity.